R Notebook
Title level 1
This is the second day of the carpentry workshop. heres another edit
Title level 2
bold
italics
Load packages
library(tidyverse)## Loading tidyverse: ggplot2
## Loading tidyverse: tibble
## Loading tidyverse: tidyr
## Loading tidyverse: readr
## Loading tidyverse: purrr
## Loading tidyverse: dplyr## Conflicts with tidy packages ----------------------------------------------## filter(): dplyr, stats
## lag(): dplyr, statslibrary(plotly)##
## Attaching package: 'plotly'## The following object is masked from 'package:ggplot2':
##
## last_plot## The following object is masked from 'package:stats':
##
## filter## The following object is masked from 'package:graphics':
##
## layoutsource("functions.R")I downloaded the file and loaded it into R
download.file("https://raw.githubusercontent.com/swcarpentry/r-novice-gapminder/gh-pages/_episodes_rmd/data/gapminder-FiveYearData.csv", destfile = "data/gapminder-FiveYearData.csv")
gapminder <- read.csv("data/gapminder-FiveYearData.csv")
head(gapminder)## country year pop continent lifeExp gdpPercap
## 1 Afghanistan 1952 8425333 Asia 28.801 779.4453
## 2 Afghanistan 1957 9240934 Asia 30.332 820.8530
## 3 Afghanistan 1962 10267083 Asia 31.997 853.1007
## 4 Afghanistan 1967 11537966 Asia 34.020 836.1971
## 5 Afghanistan 1972 13079460 Asia 36.088 739.9811
## 6 Afghanistan 1977 14880372 Asia 38.438 786.1134I wonder if rstats increases life expectancy over the years
p <- ggplot(data=gapminder,aes(x=year,y=lifeExp)) +
geom_point()
p
Let’s see the interactive version
ggplotly(p)## We recommend that you use the dev version of ggplot2 with `ggplotly()`
## Install it with: `devtools::install_github('hadley/ggplot2')`Making your own functions
If you are repeating yourself in your code, you may be able to solve that problem by making your own function!
se <- function(x){
sd(x)/sqrt(length(x))
}
cars <- c(3,4,5,6,7,10)
se(cars)## [1] 1.013794Data manipulation with dplyr
You will likely want to get subsections of your dataframe and/or calculate means of a variable for a certain subsection, dplyr is your friend!
Explored select
gapminder <- read.csv("data/gapminder-FiveYearData.csv")
year_country_gdp <- select(gapminder,year,country,gdpPercap)
year_country_gdp <- select(gapminder,-pop, -continent, -lifeExp)
names(year_country_gdp)## [1] "country" "year" "gdpPercap"Explore filter Pike (%>%) = first argument of the function (shortcut: control shift m)
year_country_gdp_euro <- gapminder %>%
filter(continent=="Europe") %>%
select(year,country,gdpPercap)Exploring the amazing group_by and summarize functions
mean_gdp_percountry <- gapminder %>%
group_by(country) %>%
summarise(mean_gdp=mean(gdpPercap),
se_gdp=se(gdpPercap))
mean_gdp_percountry## # A tibble: 142 x 3
## country mean_gdp se_gdp
## <fctr> <dbl> <dbl>
## 1 Afghanistan 802.6746 31.23550
## 2 Albania 3255.3666 344.20223
## 3 Algeria 4426.0260 378.26190
## 4 Angola 3607.1005 336.56641
## 5 Argentina 8955.5538 537.68144
## 6 Australia 19980.5956 2256.11315
## 7 Austria 20411.9163 2787.23968
## 8 Bahrain 18077.6639 1563.29518
## 9 Bangladesh 817.5588 67.86165
## 10 Belgium 19900.7581 2422.32683
## # ... with 132 more rowsChallenge: I want the mean, se and sample size of life expectancy by continent n() = easier way to find “sample size” instead of length (exists only within summarise or another fnx) you can group_by multiple variables (just by adding a comma)
mean_lifeExp_percontinent <- gapminder %>%
group_by(continent,country) %>%
summarise(mean_lifeExp=mean(lifeExp),
se_lifeExp=se(lifeExp),
length_lifeExp=n())
mean_lifeExp_percontinent## # A tibble: 142 x 5
## # Groups: continent [?]
## continent country mean_lifeExp se_lifeExp
## <fctr> <fctr> <dbl> <dbl>
## 1 Africa Algeria 59.03017 2.9849208
## 2 Africa Angola 37.88350 1.1562236
## 3 Africa Benin 48.77992 1.7691977
## 4 Africa Botswana 54.59750 1.7116922
## 5 Africa Burkina Faso 44.69400 1.9762099
## 6 Africa Burundi 44.81733 0.9165096
## 7 Africa Cameroon 48.12850 1.5784640
## 8 Africa Central African Republic 43.86692 1.3627459
## 9 Africa Chad 46.77358 1.4110376
## 10 Africa Comoros 52.38175 2.3476081
## # ... with 132 more rows, and 1 more variables: length_lifeExp <int>combining ggplot and dplyr easy tools to save = ggsave and write.cvs (“/” denotes what folder you put it in - don’t put it in your “data” folder because there is a risk of over ridding your original data)
euro_countries <- gapminder %>%
filter(continent=="Europe") %>%
ggplot(aes(x=year,y=lifeExp,color=country)) +
geom_line() +
facet_wrap(~country)
euro_countries
ggsave("euro.png")## Saving 7 x 5 in imagewrite.csv(mean_gdp_percountry, "processed/mean_gdp_percountry.cvs")Data manipulation with tidyr
make sure you use the right data frame (gapminder_wide vs. gapmidner_long)
a longer way to code this data… gap_long <- gapminder_wide %>% gather(obstype_year, obs_values, starts_with(‘pop’), starts_with(‘lifeExp’), starts_with(‘gdpPercap’) )
# command to download the 'wide' data
download.file("https://raw.githubusercontent.com/swcarpentry/r-novice-gapminder/gh-pages/data/gapminder_wide.csv", destfile = "data/gapminder_wide.csv")
gapminder_wide <- read.csv("data/gapminder_wide.csv")
gap_long <- gapminder_wide %>%
gather(obstype_year,
obs_values,
3:38)
head(gap_long)## continent country obstype_year obs_values
## 1 Africa Algeria gdpPercap_1952 2449.0082
## 2 Africa Angola gdpPercap_1952 3520.6103
## 3 Africa Benin gdpPercap_1952 1062.7522
## 4 Africa Botswana gdpPercap_1952 851.2411
## 5 Africa Burkina Faso gdpPercap_1952 543.2552
## 6 Africa Burundi gdpPercap_1952 339.2965seperate the obs_type column
gap_normal <- gap_long %>%
separate(obstype_year,into=c("obs_type", "obs_year"
),sep="_") %>%
spread(obs_type,obs_values)
head(gap_normal)## continent country obs_year gdpPercap lifeExp pop
## 1 Africa Algeria 1952 2449.008 43.077 9279525
## 2 Africa Algeria 1957 3013.976 45.685 10270856
## 3 Africa Algeria 1962 2550.817 48.303 11000948
## 4 Africa Algeria 1967 3246.992 51.407 12760499
## 5 Africa Algeria 1972 4182.664 54.518 14760787
## 6 Africa Algeria 1977 4910.417 58.014 17152804all.equal(gapminder,gap_normal)## [1] "Names: 5 string mismatches"
## [2] "Component 1: Attributes: < Component \"levels\": Lengths (142, 5) differ (string compare on first 5) >"
## [3] "Component 1: Attributes: < Component \"levels\": 5 string mismatches >"
## [4] "Component 1: 1704 string mismatches"
## [5] "Component 2: Attributes: < target is NULL, current is list >"
## [6] "Component 2: target is numeric, current is factor"
## [7] "Component 3: Modes: numeric, character"
## [8] "Component 3: target is numeric, current is character"
## [9] "Component 4: 'current' is not a factor"
## [10] "Component \"lifeExp\": Mean relative difference: 0.203822"
## [11] "Component 6: Mean relative difference: 4101.546"R likes to have ‘long’ format data where every row is an observation and you have a single column for ‘observations’ the others serve to identify that observation. (exceptions apply when you have multiple types of observations) To switch back and forth from ‘wide’ (how we typically enter data in a spreadsheet) to ‘long’ use tidyr